It's Not What the Image Shows: Irrelevant Context Destabilises VLM Judges Without Informing Them
Paper • 2609.37863 • Published • 38 •
3
ResearchStudio Bot
researchstudio-bot
AI & ML interests
Official ResearchStudio bot for sharing human-reviewed interactive Paper Reels.
Recent Activity
commentedon a paper about 12 hours ago
It's Not What the Image Shows: Irrelevant Context Destabilises VLM Judges Without Informing Them commentedon a paper about 13 hours ago
Unmask the State: When Does State Adaptation Matter for Masked Diffusion Language Models commentedon a paper about 14 hours ago
TimeEvo: Failure-Driven Self-Evolution of a Time Series AgentOrganizations
None yet
commented a
paper about 12 hours ago
commented a
paper about 13 hours ago
commented a
paper about 14 hours ago
updated a
bucket about 16 hours ago
commented 12
papers 4 days ago
In-Flight KV Cache with Clean Anchors for Faster Autoregressive Video Diffusion
Paper • 2609.32540 • Published • 34 •
2
Learning from Teacher Continuations at Student States
Paper • 2609.36246 • Published • 40 •
2
RSIGame: Autonomous Agentic Game Development with Recursive Self-improvement
Paper • 2609.39045 • Published • 85 •
2
More Choices, Fewer Decisions: Ordinal-Scale Bias in JEV-like Direct-Decision Models
Paper • 2609.38827 • Published • 57 •
3
Skill2Env: Capability-Oriented Environment Synthesis from Skills for General Agents
Paper • 2609.33772 • Published • 35 •
7
EVOKE: Eliciting World Knowledge in Agents for Transferable Decision-Making
Paper • 2609.38334 • Published • 74 •
3
AREX-2: Advancing Self-Improving Agents through Long-Horizon Reflective Tasks
Paper • 2609.38288 • Published • 134 •
3
UniEvo-VL: An On-policy Self-Distillation Training Recipe for Multimodal Model Self-improvement
Paper • 2609.38721 • Published • 280 •
3
WorldAuditBench: Interactive 3D World Auditing with Multimodal Agents
Paper • 2609.40325 • Published • 97 •
3
The Teacher Is a Direction, Not a Destination: Extrapolating RL-Induced Representation Residuals in On-Policy Distillation
Paper • 2609.36484 • Published • 539 •
3
APM-Bench: Benchmarking Cross-session Persistent Memory for Egocentric Streaming Video Assistants
Paper • 2609.37559 • Published • 46 •
3
Selecting Diverse SFT Traces Improves Post-RL Generalization
Paper • 2609.33780 • Published • 38 •
2
commented 2
papers 7 days ago
RayOrch: Programming and Executing Lineage-Controlled Multi-Grain Dataflows for Foundation-Model Data Preparation
Paper • 2609.18703 • Published • 55 •
3
FuseReg: Regularizing Layer Fusion Mitigates the Reconstruction-Generation Gap in Representation Autoencoders
Paper • 2609.31620 • Published • 159 •
3
commented 2
papers 11 days ago